Ecological observations of pelagic bacterial and archaeal communities in the Atlantic-Arctic boundary zone

(1)

Ecological observations of pelagic

bacterial and archaeal communities in

the Atlantic-Arctic boundary zone

D

ISSERTATION

ZUR

E

RLANGUNG DES

D

OKTORGRADES DER

N

ATURWISSENSCHAFTEN

- D

R

.

RER

.

NAT

.

-DEM

F

ACHBEREICH

G

EOWISSENSCHAFTEN DER

U

NIVERSITÄT

B

REMEN

VORGELEGT VON

Eduard Fadeev

(2)

Die vorliegende Doktorarbeit wurde im Rahmen des Pro-gramms International Max Planck Research School of Marine Microbiology (MarMic) in der Zeit von November 2015 bis Oktober 2018 in der HGF MPG Brückengruppe für Tiefsee-Ökologie und Technologie am Alfred-Wegener-Institut Helmholtz-Zentrum für Polar- und Meeresforschung und dem Max-Planck-Institut für Marine Mikrobiologie angefertigt.

Gutachter: Prof. Dr. Antje Boetius Gutachter: Dr. Daniel Sher Prüfer: PD Dr. Bernhard Fuchs Prüfer: Prof. Dr. Kai-Uwe Hinrichs Prüfer: Dr. Morten H Iversen

(3)

Summary

The global climate change has an unprecedented impact on the Arctic Ocean, resulting in warm-ing of the Arctic surface air at much faster rates than the global average. The warmwarm-ing temper-atures lead to constantly declining Arctic sea ice cover, which reached in September 2018 the sixth lowest summertime minimum extent in the satellite record (since the late 1970s). Shrink-ing sea ice has a strong impact on the entire Arctic marine ecosystem, through alterations of the primary production, grazers communities, and subsequently the biological carbon pump. Cur-rent predictions of entirely sea-ice free summers in the Arctic Ocean already in the second half of this century urges the need to understand the ongoing oceanographic and biological processes in order to predict how the Arctic ecosystem will respond to further environmental changes. The differentiation between natural temporal ecosystem variability and anthropogenically-induced impact of the climate change requires long-term observations. The Ocean Observing System FRAM (FRontiers in Arctic marine Monitoring), which was established in 2014, is an Arc-tic long-term observatory for investigating the impact of changing ocean properties and sea ice conditions of the Arctic Ocean on its marine ecosystem. The starting point for the FRAM project was the already existing long-term observatory HAUSGARTEN, situated in the main gateway between the Arctic and the Atlantic Oceans - the Fram Strait. To date, despite their impor-tance for the biogeochemical cycling, very little is known regarding the diversity and function of microbial communities in the Arctic Ocean in general, and speciﬁcally in the Fram Strait. In the framework of FRAM, a Molecular Observatory was established, for conducting standardized molecular-based high-resolution observations of the Arctic microbial communities.

This thesis was conducted as part of the FRAM Molecular Observatory, and as part of the estab-lishment process of the observatory it contributes to the methodological and procedural stan-dardization required for long-term microbial observations. This thesis provides a ﬁrst compre-hensive overview of currently existing long-term microbial observatories around the world, it provides guidelines for initial steps towards establishing a community network between them, and stresses the urgent need in community efforts towards methods standardization. Further-more, as part of the methods standardization for long-term microbial observations, this thesis includes a performance comparison between two, broadly used in microbial oceanography, 16S rRNA gene primer sets.

The main focus of the thesis is on the ecology of pelagic bacterial and archaeal communities in the Fram Strait. Its overall objective was to investigate the distribution of these communities in the Fram Strait, and to identify environmental drivers of their diversity. The observations of this thesis reveal that sea ice has a strong impact on the development of the seasonal phytoplankton

(4)

bloom during the summer. As a result, sea ice conditions are affecting the bacterial diversity in surface water, and are leading to a distinct community in sea-ice free and sea-ice covered regions of the Fram Strait. However, the impact of the sea ice is not limited to the surface ocean, as it also heavily affects the vertical export of aggregated organic matter to the deep ocean. The results of this thesis also show that aggregates formed under the sea ice sink faster, and by that provide a stronger vector for transport of bacterial and archaeal taxa to the deep ocean, compared to ice-free waters.

Altogether, this thesis contributes to the baseline knowledge needed for further long-term ob-servations of pelagic microbial communities in the Arctic marine ecosystem. Furthermore, it provides an important insight into the strong impact of the sea ice on bacterial and archaeal communities throughout the entire water column, underlining the potential impact of further environmental changes on the Arctic Ocean in the light of prevalent global warming and climate change.

(5)

Zusammenfassung

Der Einﬂuss des globalen Klimawandels ist an keinem Platz der Welt so deutlich sichtbar wie in dem Arktischen Ozean. Dies lässt sich besonders an der Erwärmung der Luft über der Arktis messen, die sich hier wesentlich schneller erwärmt als im globalen Durchschnitt. Diese warmen Temperaturen führten in den letzten Jahren zu einer konstant schmelzenden Eisdecke. Dieses Jahr im September wurde seit Anbeginn der Messungen (1970) die sechstniedrigste Eisstärke auf dem Arktischen Ozean gemessen.

Die Abschmelzung der Eisdecke auf dem Arktischen Ozean hat extreme Folgen für das gesamte marine Ökosystem, da es zu Veränderungen in der Primärproduktion und der Zooplank-tonzusammensetzung kommt, was schlussendlich einen Effekt auf die biologische Kohlenstoff-pumpe besitzt. Man geht momentan davon aus, dass bereits in der zweiten Hälfte dieses Jahrhunderts die Arktis im Sommer komplett eisfrei sein wird. Diese starken Veränderungen der Umwelt werden die ozeanografischen und biologischen Prozesse zwangsläufig beeinflussen. Daher ist es unabdingbar zu verstehen, wie sich die Klimaerwärmung bereits heute auf unsere Ökosysteme auswirkt, um zukünftige Veränderungen besser prognostizieren zu können. Um den Einfluss des Klimawandels im Grundsatz zu verstehen und um unterscheiden zu kön-nen zwischen natürlicher, zeitlicher Variabilität des Ökosystems und anthropogekön-nen Einflüssen, braucht es Langzeitstudien. Das Ozean-Beobachtungssystem FRAM (FRontiers in Arctic marine Monitoring) ermöglicht genau diese Langzeitbeobachtungen, welche sich auf den Einfluss der Veränderungen des Ozeans und der Eisverhältnisse auf das marine Ökosystem der Arktis konzen-trieren. Den Grundstein für das FRAM Projekt legte das bereits seit langem bestehende Obser-vatorium HAUSGARTEN, welches sich in dem Hauptzugang vom Arktischen zum Atlantischen Ozean befindet- der Framstraße.

Bis heute weiß man sehr wenig über die generelle Vielfalt und die Funktion mikrobiologischen Lebens in der Arktis, obwohl dies für das Verständnis des biogeochemikalischen Kreislaufs ins-besondere in der Framstraße unbedingt notwendig ist. Im Rahmen des FRAM Projektes wurde daher ein Observatorium aufgebaut, welches es ermöglicht, genau diese Prozesse durch standar-disierte molekulare Methoden hoch auﬂösend zu beobachten und besser zu verstehen wie die arktischen mikrobiologischen Gemeinschaften funktionieren.

Die Forschungsarbeit auf der diese Doktorarbeit basiert, wurde im Rahmen des mikrobiellen FRAM Observatoriums ermöglicht und trägt zu einer Methodenentwicklung und Prozessstan-dardisierung bei, die unbedingt für mikrobiologische Langzeitbeobachtungen dieser Region benötigt werden. Diese Arbeit gibt zudem den ersten allumfassenden Überblick aller derzeit ex-istierenden mikrobiologischen Langzeitobservatorien, sowie Handlungsempfehlungen für erste

(6)

Schritte um ein Kommunikationsnetzwerk zwischen ihnen aufzubauen und betont die dringende Notwendigkeit Methoden zu standardisieren. Aus diesem Grunde werden in dieser Arbeit als er-ster Schritt Richtung Methodenentwicklung die zwei meist genutzten 16S rRNA Genprimersets miteinander verglichen.

Der Schwerpunkt dieser Doktorarbeit liegt auf der Ökologie der pelagischen bakteriellen und ar-chaeellen Gemeinschaften in der Framstraße mit dem Ziel, die Verteilung dieser Gemeinschaften zu untersuchen und die Haupteinflüsse der Umwelt auf ihre Artenvielfalt zu identifizieren. Die Arbeit kommt zu zwei Hauptergebnissen: zum einen hat die Ausbreitung von Meereseis einen großen Einfluss auf die Entwicklung der saisonalen Planktonblüte im Sommer und die Vielfalt der Bakterien im Oberflächenwasser, was zu sich deutlich unterscheidenden Diversität der Gemein-schaften in vereisten und nicht vereisten Regionen auf allen Wasserebenen der Framstraße führt. Dies bedeutet, dass nicht nur das Oberflächenwasser von den Veränderungen beeinflusst ist, son-dern genauso auch der vertikale Export aggregierter organischer Materie in die Tiefsee. Außer-dem zeigt diese Arbeit, dass Aggregate, welche unter Außer-dem Meereis entstanden, schneller sinken und einen schnelleren Transportvektor bakterieller und archaeeller Taxa in die Tiefsee darstellen im Vergleich zu Aggregaten, die sich in eisfreien Gebieten bildeten.

Insgesamt liefert die Doktorarbeit einen entscheidenden Beitrag zum Aufbau des Basiswissens zur Langzeitbeobachtungen von pelagischen mikrobiellen Gemeinschaften des arktischen mari-nen Ökosystems, welches für zukünftige Forschung unabdingbar ist. Des Weiteren liefert sie wichtige Erkenntnisse über den Einﬂuss des Meereseis auf bakterielle und archaeelle Gemein-schaften innerhalb der gesamten Wassersäule, welches den potentiellen Einﬂuss von weiteren Veränderungen der Umwelt auf den Arktischen Ozean im Hinblick auf die globale Erwärmung und den Klimawandel weiter unterstreicht.

(7)

List of Figures

1.1 Bathymetry and hydrography of the Arctic Ocean . . . 3

1.2 Map of FRAM infrastructure in the Arctic Ocean and the Fram Strait . . . 15

2.1 Map of marine long term ecological time series sites with microbiome variables monitored. . . 34

3.1 Sampling coverage of bacterial communities used for primer sets comparison. . . 42

3.2 Mean sequence proportion of taxonomic classes in each habitat . . . 43

3.3 Differential abundance of bacterial families between the primer sets. . . 45

3.4 Relative abundance of bacterial taxa using 16S rRNA gene and CARD-FISH . . . . 46

3.S1 Sequence retainment throughout the bioinformatic workﬂow . . . 50

3.S2 Linear correlation in sequence proportions between the primer sets . . . 51

3.S3 Correlation of 16S rRNA and CARD-FISH bacterial relative abundance . . . 52

4.1 Oceanographic overview of Fram Strait during June 2014 . . . 60

4.2 Modeled weekly surface chl a concentration in Fram Strait during June 2014 . . . 61

4.3 Comparison of bacterial community composition between the different regions and fractions . . . 62

4.4 Sequence proportion overview of overlapping bacterial OTUs. . . 63

4.5 Enriched bacterial families between the regions . . . 65

4.6 Sequence proportion overview of overlapping microbial eukaryote OTUs. . . 66

4.7 Bacterial community characteristics across the Fram Strait. . . 67

4.8 RDA ordination of bacterial community composition . . . 68

4.9 Overview of edge counts for selected taxonomic groups in each network . . . 69

4.10 Sub-networks of the FL and PA fractions in the chl. a max. . . 70

4.S1 Regional separation of Fram Strait based on in situ biogeochemical parameters. . 80

4.S2 Rarefaction analysis of bacterial and eukaryotic communities . . . 81

4.S3 Differences in bacterial community composition between the regions . . . 82

4.S4 Venn diagram of shared OTU . . . 83

4.S5 Enriched microbial eukaryotic taxonomic groups between the regions. . . 84

4.S6 PCoA of microbial eukaryote community composition. . . 85

4.S7 Venn diagram of shared OTU . . . 85

4.S8 Monthly mean of surface chlorophyll a in FESOM-REcoM2 . . . 88

4.S9 Comparison of modeled and in situ measured parameters across the Fram Strait . 88 5.1 Oceanographic overview of Fram Strait during July 2016 . . . 98

5.3 Exemplary light microscopy images of marine aggregates in Fram Strait . . . 101

5.4 Sinking trajectories of particles. . . 102

5.5 Microbial community composition throughout the water column. . . 104

5.6 Free-living and particle-associated community dynamics throughout the water column of Fram Strait . . . 105

5.7 Differences in PA community composition between the the distinct water layers . 107 5.8 SourceTracker estimates of water mass contribution. . . 109

5.S1 Sample-size-based rarefaction curves of microbial communities. . . 116

(8)

5.S2 Chlorophyll a concentrations along Fram Strait during July 2016. . . 116

5.S3 Vertical changes in communitiy richness and diversity. . . 117

5.S4 Dissimilarities between microbial fractions throughout the water column. . . 117

6.1 Comparison of bacterial and archaeal communities between the central Arctic

Ocean and the Fram Strait. . . 124

(9)

List of Tables

2.1 Details of samples used in Polar Front study. . . 22

3.S1 Details of samples used for primers comparison . . . 53

3.S2 Details of CARD-FISH probes . . . 55

4.1 Comparison of nutrient consumption, phytoplankton biomass and productivity . . 61

4.S1 Overview of sampled stations during RV Polarstern expedition PS85. . . 89

4.S2 Bacterial cell abundance and activity . . . 92

4.S3 Properties of the co-occurrence networks . . . 93

5.1 Characteristics of particles in ice-free and ice-covered regions of Fram Strait. . . . 100

5.S1 Settings of particle tracking experiments. . . 115

5.S2 Overview of sampled stations during RV Polarstern expedition PS99.2 . . . 118

5.S3 Microbial community alpha diversity estimations. . . 120

(10)

List of acronyms

AW Atlantic Water

AtlantOS Atlantic Ocean Observing Systems

CARD-FISH CAtalyzed Reporter Deposition-Fluorescence In Situ Hybridization CTD conductivity temperature depth

DOM dissolved organic matter DNA deoxyribonucleic acid

daOTU differentially abundant operational taxonomic unit EBDW Eurasian Basin Deep Water

EGC East Greenland Current EMP Earth Microbiome Project EOV Essential Ocean Variables

EPS extracellular polymeric substances FL free-living

FISH Fluorescence In Situ Hybridization FRAM FRontiers in Arctic marine Monitoring

HNA high nucleic acids LNA low nucleic acids

LTER Long-term ecological research MAW mixed Atlantic Water

MSC Marine Snow Catcher MolObs Molecular Observatory

MI microbial indicator

OTU operational taxonomic unit OM organic matter

PA particle-associated PCR polymerase chain reaction POM particulate organic matter

PPS PhytoPlankton Sampler PW Polar water

(11)

LIST OF TABLES xi

RAS Remote Access water Sampler RNA ribonucleic acid

rRNA ribosomal RNA

TEP transparent exopolymer particles UVP Underwater Vision Proﬁler WSC West Spitsbergen Current

(12)

Introduction

1.1 The Arctic Ocean

The Arctic Ocean is the smallest ocean on Earth, making up ~4% of the global ocean area

and only ~1% of its volume (Jakobsson, 2002). The Arctic Ocean has a unique circumpolar

oceanography, characterized by strong seasonal cycles in temperature, solar irradiation and

sea-sonal formation of sea ice (Johannessen et al.,1994). Unlike the Southern Ocean, the Arctic

Ocean is almost completely enclosed by land and is classiﬁed as a mediterranean sea,

compris-ing 35% of the world coastline (Tomczak and Godfrey,2013). Furthermore, about 50% of the

Arctic Ocean surface area is comprised of continental shelves, which are considered to be the

most fertile regions of the Arctic Ocean (Carmack and Wassmann,2006). Its central basin is

separated by the subsurface Lomonosov Ridge into the Amerasian Basin, which connects to the

North Paciﬁc Ocean, and the Eurasian Basin, which connects the North Atlantic Ocean (

Fig-ure 1.1;Beszczynska-Möller et al.,2011). Despite its small size, the Arctic Ocean has important

functions in the global ocean circulation. It absorbs the heat from the north Atlantic thermo-haline circulation, and produces the colder and denser waters that form the deep layers of the

North Atlantic (Aagaard et al.,1985). It also provides an oceanic connection between the North

Atlantic and North Paciﬁc Oceans that has an important role in the nutrient ﬂuxes between these

oceans (Torres-Valdés et al.,2013).

1.1.1 Physical oceanography

The Arctic Ocean exchanges water with both the Paciﬁc and the Atlantic Oceans through two

main oceanic gateways, Bering Strait and Fram Strait (Beszczynska-Möller et al.,2011). The low

salinity Paciﬁc waters are entering the Arctic Ocean through the shallow (50 m) Bering Strait. These waters are characterized by high silicate and phosphate concentrations, providing a large

source of nutrients for the Arctic Ocean (Torres-Valdés et al.,2013). However, they comprise

only a small fraction of the total inﬂow to the Arctic Ocean. In contrast, Atlantic inﬂow through

the Fram Strait is roughly 10 times larger than the Paciﬁc inﬂow (Beszczynska-Möller et al.,

2011). The Atlantic waters (AW) are more saline and have higher nitrate to phosphate (N:P)

ratio, compared to the Paciﬁc waters. At depths of more than 1000 m, the Arctic Deep Water 1

(16)

layer begins (Smith Jr,2013). This very dense homogeneous layer comprises the Arctic Ocean

deep waters, and reaches the seaﬂoor (Figure 1.1).

In general, the surface waters of the Arctic Ocean are stratiﬁed by salinity, with relatively weak

vertical mixing (Carmack,2007). In both basins, the water column is characterized by a low

salinity polar mixed layer at the surface (10 m depth), separated by a cold halocline from the

layers below (Rudels et al.,1996). Being relatively fresher, the Paciﬁc waters lie higher in the

water column than Atlantic waters (AW), and provide part of the strong stratiﬁcation of the Arctic water column. This strong water column stability isolates the surface waters from the

heat of the Atlantic and the Paciﬁc inﬂows, and allows formation of sea ice in winter (Rudels,

2012).

The most deﬁning characteristic of the Arctic Ocean is the sea ice, that covers in winter the

entire Arctic Ocean (Thomas,2017). In the dark autumn and winter the temperature of surface

waters drops below the freezing point. During that time the Arctic sea ice is formed, reaching its

maximum seasonal coverage in March (Polyak et al.,2010). The sea ice formation process results

in brine rejection that increases the salinity of the underlying waters. Weakened stratiﬁcation of the water column initiates deep vertical mixing, which forms the Arctic bottom water and "ﬁlls

up" the nutrients budget in the surface (Korhonen et al.,2013). In summer, as a result of solar

radiation and warmer incoming waters, the sea ice melts, and reaches its seasonal minimum in September. The melting process releases freshwater and strengthens the stratiﬁcation of the

water column once again (Korhonen et al.,2013).

The sea ice buffers interactions (e.g., heat exchange) between the atmosphere and the ocean,

and governs light availability in the underlying water column (Perovich and Polashenski,2012).

Furthermore, it prevents surface and internal waves, and isolates the water surface from winds

(Thomas and Dieckmann,2003). As such, it is affected by both the winds above and the

wa-ter currents below, and is carried with the transpolar drift towards the Fram Strait (Figure 1.2

Pﬁrman et al.,1997).

The concentration and thickness of the sea ice are a result of thermodynamic processes and

provide an important evidence for the global climate change (Gao et al.,2015;Budikova,2009).

The shrinking sea ice extent results in surface warming twice as fast as the global average (Sun

et al.,2016;Dobricic et al.,2016). Warmer surface waters are weakening the cold halocline layer, allowing stronger vertical mixing and upward AW heat ﬂux, which further ampliﬁes the

sea-ice loss (Polyakov et al.,2017). According to current model projections the Arctic Ocean

may experience sea-ice free summers already by the second half of this century (Overland and

Wang,2013;Wang and Overland, 2015). Such fundamental change of the Arctic Ocean will

lead to strong alterations of the marine ecosystem (Wassmann and Reigstad,2011;Arrigo et al.,

2008).

1.1.2 Primary production

On a global scale, one of the key biological processes that occurs in the ocean is primary

produc-tion by photosynthetic organisms (Falkowski et al.,1998;Field,1998). This process is fueled

by solar radiation and provides the basis of the marine food web, as it converts inorganic car-bon sources (such as carcar-bon dioxide) into bioavailable organic carcar-bon. Primary production by

(17)

1.1. THE ARCTIC OCEAN 3

Figure 1.1: (A) Map of the Arctic Ocean seaoor features. Modied from: www.ngdc.noaa.gov/mgg/bathymetry/arctic/currentmap.html. The red line represent a hy-pothetical transect from Bering Strait to Fram Strait. (B) Schematic vertical distribution of the major water masses across the Arctic Ocean based on the hypothetical transect from Bering Strait to Fram Strait. Adapted fromGonçalves-Araujo (2016)

(18)

photosynthetic organisms is often limited by the availability of additional nutrients, such as

ni-trogen and phosphorous, which are required for the carbon ﬁxation (Howarth,1988). In the

Arctic Ocean, primary production is also constrained by the availability of solar radiation, which

is governed by the presence of sea ice (Popova et al.,2012;Arrigo,2014). Furthermore, due to

the low angle of the sun, overall the Arctic Ocean receives less solar radiation and has a lower

annual primary production compared with other oceanic regions (Lee et al.,2015).

The primary production rates in the Arctic Ocean differ strongly between the deep central basin

(i.e., central Arctic Ocean) and the continental shelves (Carmack and Wassmann,2006). The

broad shelf areas of the Arctic Ocean are seasonal sea-ice zones which receive 11% of the global

river runoff and can sustain high primary productivity (Aagaard and Carmack,1989;Tremblay

and Gagnon,2009;Carmack and Wassmann,2006). The inﬂow shelves, through which nutrient

rich sub-Arctic waters are entering the Arctic Ocean, are considered to be by far the most fertile

regions in the Arctic Ocean (Wassmann,2015). The central basin, on the other hand, due to high

sea-ice coverage and low nutrient availability, is considered to be signiﬁcantly less productive

(Tremblay et al.,2015;Sakshaug,2004).

Unlike the sub-Arctic regions, the Arctic Ocean is shaped by extreme seasonality with three months a year of complete darkness (winter) and three months of permanent daylight (summer). During the dark winter time, as a result of vertical mixing and lack of active primary production,

the nutrient level in the surface waters reaches its annual maximum (Codispoti et al.,2013).

In spring, with the increase in light availability the seasonal phytoplankton bloom begins (Leu

et al.,2011). There are two main sources for primary production in the Arctic Ocean, sea-ice algae and pelagic unicellular phytoplankton. Due to differences in light sensitivity, these two

groups of primary producers differ in the timing of their bloom (Terrado et al.,2013). The

sea-ice algae are adapted to lower light conditions, and therefore able to start growing earlier in

the spring (Hancke et al.,2018;Arrigo,2014). The total production of sea-ice algae is highly

variable and depending on the sea-ice situation, however it is estimated between 5 and 10 g

C m-2 yr-1 (Gosselin et al., 1997;Leu et al.,2011). The sea-ice algae contribute more than

50% of the primary production in the central Arctic Ocean, and only up to 25% of the primary

production in the Arctic shelf regions (Gosselin et al.,1997;Legendre et al.,1992). With the

increasing availability of solar radiation and due to retreating sea ice, the pelagic phytoplankton

bloom usually occurs on the sea-ice edge in short massive bloom events (Arrigo et al.,2012;

Fernández-Méndez et al.,2015), and is estimated to 12-50 g C m-2yr-1(Gosselin et al.,1997;

Leu et al.,2011).

The primary production in the Arctic Ocean is carried out by unicellular eukaryotic algae, such

as, diatoms, dinoﬂagellates and haptophytes (Poulin et al.,2011;Terrado et al.,2013). Overall,

diatoms are considered to be the main contributors to primary production in the Arctic Ocean

(Gosselin et al., 1997). Diatoms are an extremely diverse taxonomic group with more than

10,000 species (Smetacek,2000), characterized by a siliciﬁed cell wall which provides

mechan-ical protection from grazers (Hamm et al.,2003). They are unicellular organisms which may

occur both as solitary cells and in colonies, ranging in size over several orders of magnitude

(Smetacek,2000). They are found in a wide range of freshwater and marine environments, and

in the Arctic Ocean are present both in the water column and associated with sea ice (Arrigo,

2014). The most common pelagic diatom species in the Arctic Ocean belong to the lineages

(19)

1.1. THE ARCTIC OCEAN 5

the sea ice are pennate diatoms, such as, Nitzschia and Navicula (Quillfeldt,2005). In addition,

the centric diatom Melosira arctica is known to form long ﬁlaments attached to the bottom of

the sea ice (Boetius et al.,2013).

1.1.3 Vertical export of organic matter

The oceans are the largest reservoir of carbon in the biosphere, on time scales of hundreds to thousands of years, and therefore they play a central role in the regulation of the increasing

atmospheric CO2 concentrations (Takahashi et al.,2002). The Arctic Ocean is responsible for

an uptake of 66-199 Tg C yr-1, contributing up to 14% to the global uptake of CO2 (Bates and

Mathis,2009). Part of the sequestered atmospheric CO2is due to production of organic carbon by

the primary producers in the surface ocean. The vast majority of the produced organic carbon is respired, by heterotrophic organisms, back to carbon dioxide already in the surface ocean

(Ducklow et al.,2001). Nevertheless, up to 30% of it is exported to the deep ocean by sinking

particles of organic matter (e.g., decaying phytoplankton and fecal pellets; Turner, 2002), a

process also termed "the biological pump” (de La Rocha and Passow,2003;De La Rocha and

Passow,2007). The sinking particles of organic matter (OM) which escape the photic layer of

the ocean, provide the main source of food for the deep ocean biology (Ducklow et al.,2001).

Overall, less than 5% of the produced OM in the surface ocean are eventually reaching the

seaﬂoor (Jørgensen and Boetius,2007).

The magnitude of the vertical export is greatly dependent on the presence of zooplankton, and

microbial activity (i.e., the "microbial loop" - further discussed insection 1.2). In temperate and

tropical latitudes the cycles of the phytoplankton bloom, and the zooplankton grazing, are vary-ing within narrow limits. The relatively small community ﬂuctuations allow strong retainment

and large recycling of the produced OM in the upper part of the water column (Rivkin et al.,

1996;Calbert and Landry, Michael,2004). In contrast, in the Arctic Ocean, the link between

phytoplankton and grazers is less pronounced (Klein et al.,2002). This results in an episodic

vertical export of ungrazed OM in the beginning of the seasonal bloom (Wassmann et al.,2004,

1996), especially on the Arctic Ocean shelves (Carmack and Wassmann,2006). With

increas-ing grazincreas-ing activity throughout the season, the vertical export is declinincreas-ing and the ecosystem is shifting from an export food chain towards a retention food chain, at the end of the productive

season (Wassmànn,1997).

1.1.4 Climate change impacts

Global climate change is enhanced in high latitudes, resulting in a warming of the Arctic surface

air at much faster rates than the global average (Dobricic et al.,2016;Woodgate et al.,2012).

Combined with the increase of Atlantic heat ﬂux into the Arctic Ocean (also termed

"Atlantiﬁ-cation";Polyakov et al.,2017), this causes strong reduction in sea-ice coverage and thickness

(Peng and Meier,2017;Kwok and Rothrock,2009;Notz and Stroeve,2016), at unprecedented

rates over at least the last few thousand years (Polyak et al.,2010;Kinnard et al.,2011). With

continuous anthropogenic release of CO2, current projections suggest that the Arctic Ocean may

experience sea-ice free summers already in the second half of this century (Overland and Wang,

(20)

(Steele and Dickinson,2016) and fresher (Carmack et al.,2016). All these phenomena are

inter-connected in a positive feedback process termed "Arctic ampliﬁcation" (Serreze and Barry,2011).

The decreasing coverage of sea ice, its thinning and early melting, results in higher light

trans-mission into the water column (Arrigo et al.,2008). Furthermore, increasing atmospheric CO2

concentrations enhance the absorption of CO2in the seawater, causing a more pronounced

acid-iﬁcation than in any other ocean (Steinacher et al.,2009;Bates and Mathis,2009;Popova et al.,

2014). These abiotic changes are clearly impacting the primary production under the sea ice

and in the water column (Arrigo and van Dijken,2015). Based on the predicted changes, there

are environmental factors that may enhance the primary production, such as, higher light avail-ability due to sea ice melt, and higher nutrient availavail-ability on the Arctic shelves, as a result of

stronger discharge from rivers (Arrigo et al.,2008). However, on the other hand, there are also

environmental changes which may diminish the primary production, such as, stronger water col-umn stratiﬁcation as a result of sea ice melt, and lower light availability due to higher cloudiness

(warmer temperatures will increase evaporation and cloud formation; Bélanger et al., 2013).

Thus, the direction, and the magnitude, of climate change impact on the total primary produc-tion in the Arctic Ocean is still heavily debated.

Not only a change in total primary production is expected. Numerous evidence for shifts in composition of the Arctic phytoplankton community towards very small (<2 μm diameter)

phy-toplankton groups, such as, Prasinophytes (Degerlund and Eilertsen,2010;Metﬁes et al.,2016;

Li et al.,2009). These small nanoﬂagellates are considered to be one of the main phytoplankton

taxa impacting the biogeochemical cycles on a global scale (Schoemann et al.,2005). Unlike

diatoms which produce heavy silicate rich cells, the Prasinophytes form almost buoyant

gelati-nous colonies with very low sinking rates (Smetacek and Nicol,2005;Wolf et al.,2016). Longer

retainment in the surface ocean allows higher recycling of the OM in the upper water column. Therefore, integrated estimates suggest that Prasinophytes contribute less than 5% to the vertical

export in the Arctic Ocean (Reigstad and Wassmann,2007). Thus, further shifts in the Arctic

phytoplankton community from diatom- to ﬂagellate- dominated communities will strongly im-pact the vertical export of OM to the deep ocean.

As primary production is the basis of the food web, its further alterations are most likely to

cas-cade through the entire Arctic ecosystem (Leu et al., 2011;Sakshaug,2004). There are well

documented ecological changes in larger organisms (Post et al.,2009), such as, stronger

mis-match between the phytoplankton bloom and the reproduction cycle of Arctic copepods (e.g.,

Leu et al., 2011;Søreide et al.,2010;Weydmann et al., 2012), as well as, further migration

northwards of the Atlantic cod (e.g.,Drinkwater,2005;McBride et al.,2014;Hollowed et al.,

2013;Ingvaldsen et al.,2017). On contrary, the ongoing changes in the microbial communities,

and speciﬁcally the bacterial and archaeal communities, remain largely understudied. Never-theless, the existing evidence suggests that these communities are also likely to be affected by

the changing conditions (Piontek et al.,2015;Brussaard et al.,2013;Sala et al.,2010;Kritzberg

(21)

1.2. MICROBIAL ECOLOGY IN THE ARCTIC OCEAN 7

1.2 Ecology of bacterial and archaeal communities in the

Arc-tic Ocean

Microorganisms perform key functions in the oceanic biogeochemical cycles by mediating the

ﬂuxes of matter, and energy, through the ecosystem (Azam and Malfatti,2007). A large fraction

of the primary production in the ocean becomes dissolved in the water column (i.e., dissolved organic matter - DOM) and is almost exclusively accessible to heterotrophic bacteria and archaea

(Thornton,2014;Sarmento and Gasol,2012). Heterotrophic microorganisms also utilize a

sig-niﬁcant fraction of the OM found in marine particles (i.e., particulate organic matter- POM) by

enzymatic “digestion” (Biddanda and Benner,1997;Arnosti et al.,2011). As a result, a large

fraction of the oceanic primary production is consumed by heterotrophic microorganisms, which

has a strong impact on the global elemental cycle (Azam,1998). The consumed OM is either

directly respired to CO2or converted into biomass which is then channeled into the "microbial

loop" (Azam et al.,1983).

1.2.1 Microbial heterotrophic activity

The term "microbial loop" refers to a complex microbial food web of production and

decompo-sition, based on the uptake and metabolism of DOM (Azam et al.,1983). In the global ocean,

heterotrophic bacteria are responsible for almost half of the community respiration, making

them the main heterotrophic microoganisms involved in OM turnover (Robinson,2008).

Bacte-rial productivity is the key pathway which "fuels" the ﬂux of OM through the loop, and therefore

is an important proxy for microbial activity in the water column (Ducklow,2002). The estimated

bacterial productivity, in different seasons and regions of the Arctic Ocean, ranges over two

or-ders of magnitude between 0.1 to 11μg C l-1_day-1₍_{Kritzberg et al.}_,₂₀₁₀_;_{Sherr and Sherr}_,₂₀₀₃_;

Sherr et al.,2003;Malmstrom et al.,2007;Nikrad et al.,2012;Nguyen et al.,2012;Kirchman

et al.,2009). These values are following the seasonal pattern of the primary production (Nikrad

et al.,2012;Nguyen et al.,2012), and are in the same range with other oceanic regions (Rich

et al.,1997).

An additional way to observe microbial activity in the water column is through cell density

dynamics (Ducklow,2002). Similar to the bacterial productivity measurements, the cell densities

in the Arctic Ocean show strong seasonal differences. Bacterial and archaeal cell densities in

the surface waters of the Arctic Ocean are estimated in the range of 105 cells ml-1 in winter,

and up to 108 cells ml-1 in summer (Sala et al., 2010;Sherr et al.,2003;Alonso-Sáez et al.,

2008;Kirchman et al.,2007), which is similar to cell densities in other oceanic regions (Morris

et al.,2002;Sunagawa et al.,2015). Interestingly, the proportion of Archaea in the community

increases with depth (Kirchman et al.,2007) and in winter (Alonso-Sáez et al.,2008). Although

very little is known regarding their role in the Arctic Ocean water column, these patterns suggest that Archaea may play an important role in the Arctic marine ecosystem.

(22)

1.2.2 Pelagic bacterial and archaeal diversity

In recent years, there were major efforts to survey the diversity of pelagic bacterial and archaeal

communities in surface (e.g., TARA Oceans;Sunagawa et al.,2015) and deep (e.g., Malaspina;

Salazar et al.,2016) waters of the global ocean, using molecular approaches. However, due to the

harsh climatic conditions and the logistical challenge, no pan-Arctic surveys of pelagic Bacteria and Archaea have been done. A synthesis of the few existing regional diversity studies show that bacterial and archaeal communities are shaped by the hydrography of the Arctic Ocean, with distinct communities in surface and deep waters.

Surface waters communities exhibit a strong seasonality in their composition. During summer they are dominated by phytoplankton-bloom associated bacteria, such as, Flavobacteria and

Gammaproteobacteria. In winter there is a strong increase in bacterial and archaeal diversity

(Ladau et al.,2013), and the community is dominated by oligotrophic taxonomic groups, such

as, the SAR11 clade (Alphaproteobacteria;Wilson et al.,2017). On contrary, deep waters

com-munities show relatively small seasonal variation. These comcom-munities are dominated by poorly

characterized taxonomic groups, such as, SAR202 clade and Marinimicrobia (Galand et al.,2010;

Ghiglione et al.,2012), and to some extent resemble the surface winter communities (Wilson

et al.,2017). Archaea are present throughout the entire water column, especially the

Thaumar-chaeota, and comprise a signiﬁcant fraction of the community in winter (Müller et al.,2018). Below, is a brief review of the current state of knowledge on the diversity, biogeography, and potential functions of key taxonomic groups of Arctic pelagic microbial communities:

• SAR11 clade - of the Alphaproteobacteria is considered to be the most abundant bacterial

lineage in the global ocean (Morris et al.,2002). All members of the SAR11 clade are

small, mostly free-living, aerobic chemoheterotrophs (Giovannoni, 2017). Their highly

streamlined genomes minimize their nutrient requirement, which potentially explain their

ecological success in the oligotrophic waters of the open ocean (Giovannoni et al.,2014;

Giovannoni,2017). The diversity within the SAR11 clade often described by nine ecotypes,

which were deﬁned based on genomic phylogeny and spatiotemporal distribution (Vergin

et al.,2012;Brown et al.,2012). Interestingly, the SAR11 surface-ocean 1a ecotype can be further divided into cold-water (1a.1) and warm-water (1a.3) subgroups, which have

distinct latitudinal distribution (Brown et al.,2012). In various molecular observations of

the Arctic Ocean, the SAR11 clade comprised up to 30% of the sequences in the community

and was present in the water column down to 1000 m (Alonso-Sáez et al.,2008;Ghiglione

et al.,2012;Balmonte et al.,2018;Wilson et al.,2017).

• Flavobacteria - are one of the most abundant bacterial taxa in the surface waters of the

Arctic Ocean (Wilson et al.,2017;Ghiglione et al.,2012;Balmonte et al.,2018).

Micro-scopic counts which targeted the broad ﬂavobacterial genus Polaribacter, using ﬂuorescent in-situ hybridization (FISH), revealed that this taxonomic group may reach up to 10% of

summer bacterial communities in Arctic waters (Malmstrom et al.,2007). Furthermore,

Flavobacteria were identiﬁed to prevail in sea ice, where they comprise around 20% of

the bacterial community (Rapp et al.,2018;Boetius et al.,2015;Bowman et al.,2012;

Brinkmeyer et al.,2003). Interestingly, Flavobacteria have been identiﬁed to exhibit strong

biogeographical partitioning both in the Southern Ocean and the North Atlantic Ocean. This has been suggested to be associated with niche separation between various clades

(23)

1.2. MICROBIAL ECOLOGY IN THE ARCTIC OCEAN 9

of this taxonomic group (Abell and Bowman, 2005; Gómez-Pereira et al., 2010). The

Flavobacteria can be found both free-living and associated with particles, and are one of

the main groups responding to phytoplankton blooms in high-latitudes (Alderkamp et al.,

2006;Pinhassi and Hagström,2000;Pinhassi et al.,2004;Chafee et al.,2018). They

pos-sess a large arsenal of hydrolytic enzymes which enable them to degrade and assimilate a

wide variety of organic biopolymers (Williams et al.,2013;Teeling et al.,2012), making

them important in remineralization of primary production products (Buchan et al.,2014).

• Gammaproteobacteria - are one of the most abundant bacterial taxa throughout the en-tire water column of the Arctic Ocean, accounting for up to 30% of the bacterial

commu-nity sequences (Kirchman et al.,2010;Wilson et al.,2017;Ghiglione et al.,2012;Balmonte

et al.,2018). These organisms can be found both free-living and particle-associated, they possess a large enzymatic arsenal that allow them rapid adaption to a wide range of carbon

sources (Buchan et al.,2014). In polar regions, similar to Flavobacteria, the

Gammapro-teobacteria exhibit strong variability in cell densities and community composition,

follow-ing the seasonal phytoplankton bloom (Wilson et al.,2017;Williams et al.,2012;

Alonso-Sáez et al.,2008). Seasonal metaproteomic analysis in Antarctic coastal waters revealed

that diverse lineages of Gammaproteobacteria signiﬁcantly increased their activity during summer in comparison to winter (e.g., Alteromonadales proportion increase from 1% in

winter to 13% in summer;Williams et al.,2012). Furthermore, Gammaproteobacteria are

also one of the dominant taxonomic groups in sea ice (Rapp et al.,2018;Bowman et al.,

2012;Boetius et al.,2015). It has been suggested that the ecological success of

Flavobac-teria and GammaproteobacFlavobac-teria in sea ice is related to their enzymatic potential to exploit

the high concentrations of exopolymeric substances (EPS) and DOM that are produced by

sea-ice algae (Grossmann and Dieckmann,1994;Aslam et al.,2012;Boetius et al.,2015).

• Verrucomicrobia - are a widespread minority phylum found in various marine

environ-ments (Freitas et al.,2012). Analyses of phenotypic, and genomic traits of a single

Verru-comicrobia isolate strain revealed their speciﬁc adaptation to utilization of glycopolymers

(Alonso-Sáez et al.,2015;Spring et al.,2016). Furthermore, this phylum has been

identi-ﬁed as the most active polysaccharide degrading taxa in the waters of Smeerenburgfjord

(Svalbard;Cardman et al.,2014). In general, little is known regarding their ecological role,

but there is increasing evidence for their potential importance in marine biogeochemical cycles.

• Deltaproteobacteria - have been identiﬁed as the second most abundant taxonomic group

in deep waters of the Arctic Ocean (Galand et al.,2010;Wilson et al.,2017). They have

been also found to comprise up to 10% of the sequences in surface communities during

winter (Wilson et al.,2017). Taxonomically, the majority of the Deltaproteobacteria

se-quences were associated with the SAR324 clade.Although, very little is known regarding this taxonomic clade, there are genomic evidences that link them to sulfur oxidation, as

well as, oxidation of methylated compounds (Swan et al.,2011). Thus, suggesting that

they may play an important role in chemoautotrophic processed in the Arctic Ocean. • SAR202 clade - of the class Dehalococcoidia is among the most dominant taxonomic groups

in the meso- and bathy- pelagic waters of the global ocean (Salazar et al.,2016), where

(24)

Arctic Ocean this group was one of the dominant members of deep water communities

(Galand et al.,2010;Bano and Hollibaugh,2002). However, they also comprised a

sig-niﬁcant proportion of sequences in the upper part of the water column in winter (Wilson

et al.,2017). The ecological niche occupied by SAR202 clade is still not fully understood, but there is evidence for their potential role in the remineralization of recalcitrant OM in

the deep ocean (Landry et al.,2017;Colatriano et al.,2018).

• Marinimicrobia (SAR406 clade) - were among the ﬁrst bacterial groups detected in the

deep ocean (Fuhrman et al.,1993), and are found to be abundant in meso- and

bathy-pelagic waters of the global ocean (Salazar et al.,2016). They have higher abundance at

low oxygen concentrations (Hawley et al., 2017), and they were found to be especially

abundant in oceanic oxygen minimum zones (Bertagnolli et al.,2017). This taxonomic

group comprised 9% of the 16S rRNA sequences in deep waters of the Arctic Ocean (Galand

et al.,2010), and also identiﬁed in surface waters during winter (Wilson et al., 2017). Genomic observations of this taxonomic group linked it to sulfur cycling via a polysulﬁde

reductase gene cluster (Wright et al.,2014;Allers et al.,2013), and to the nitrogen cycle

via expression of a nitrous oxide reductase (Hawley et al.,2017). Furthermore, it has been

shown in a methanogenic bioreactor study that Marinimicrobia participate in a syntrophic

interaction with metabolic partners to accomplish degradation of amino acids (Nobu et al.,

2015). Altogether, these traits suggest their potential importance in biogeochemical cycles

of the deep ocean.

• Thaumarchaeota - often addressed as the ammonia-oxidizing archaea (AOA), are the

most abundant pelagic archaeal group in both the surface and the deep ocean (Sunagawa

et al.,2015;Salazar et al.,2016). They consist of several phylogenetic clades which are distributed through different water layers of the water column, with a general increase in abundance with depth. It has been shown that in mesopelagic waters (1000 m) of the Arctic Ocean, the Thaumarchaeota may comprise up to 25% of microbial community

sequences (Wilson et al.,2017;Müller et al.,2018). In the surface waters of polar oceans

they exhibit seasonal patterns with an increase in relative abundance during winter and

a decline in summer (Alonso-Sáez et al.,2008;Grzymski et al.,2012). This seasonality

was previously linked to the photoinhibition of ammonia oxidation (Merbt et al.,2012),

or potentially to a stronger competition for nutrients with phytoplankton during summer

(Connelly et al.,2014;Kirchman et al.,2007).

1.3 Observing ecological changes in the Arctic Ocean

Since the beginning of the 20thcentury, oceanographers have recognized the need for extended

sampling periods to monitor and understand processes that occur in the ocean ecosystem. The longest record of sustained oceanographic observations is obtained by the Continuous Plankton

Recorder (CPR) which goes back to 1931 (Reid et al.,2003). In 1988, as part of the Joint Global

Ocean Flux Study (JGOFS), a new era of long-term observatories emerged, with the establish-ment of two major oceanographic time series programmes: the Hawaii Ocean Time-series in the

Paciﬁc Ocean (HOT;Karl and Church,2014) and the Bermuda Atlantic Time-series Study in the

(25)

obser-1.3. OBSERVING CHANGES IN THE ARCTIC OCEAN 11 vatories made a great contribution to our current understanding of the marine biogeochemistry,

the biological pump (Karl et al.,2001;Steinberg et al.,2001), and the key role of

microorgan-isms in them (Karl and Church,2014). Furthermore, these long-term observatories provided a

platform to observe impacts of climate change impact on the marine ecosystem (Ducklow et al.,

2009). As the value of such studies has become clear through the datasets produced at HOT

and BATS, the number of oceanographic observatories has continued to grow. To date, there are more than 50 oceanographic observatories around the world. Among them the LTER observa-tory HAUSGARTEN in the Fram Strait, which is the only open ocean long-term observaobserva-tory in

the Arctic Ocean (Soltwedel et al.,2005).

1.3.1 The Fram Strait and the LTER observatory HAUSGARTEN

The 450 km wide Fram Strait separates Northeast Greenland from the Svalbard Archipelago,

and is the only deep gateway to the Arctic Ocean (sill depth of ~2600 m;Hop et al.,2006).

The exchange of water in the Fram Strait occurs in both directions, by two major opposing current systems, which generate distinct physical and chemical conditions between the eastern

and western parts of the Strait (Figure 1.2).

The Atlantic inﬂow occurs through the West Spitsbergen Current (WSC) that ﬂows above the eastern shelf slope of Fram Strait. It carries the relatively warm, and saline AW northward into

the Central Arctic (Beszczynska-Moller et al.,2012). Oceanographic time-series of the WSC

reveals a positive linear trend of temperature, with an annual increase of 0.06°C in the upper

400 m of the water column, and an increase of 0.015°C at depth of 1000 m (Walczowski et al.,

2017). The Atlantic inﬂow through the Strait provides the largest input of heat into the Arctic

Ocean, sufﬁcient to melt its entire sea-ice cover (Østerhus et al.,2005). Thus, the mass and the

heat exchange through the Fram Strait have a strong impact on the entire Arctic region. In the western Fram Strait, the East Greenland Current (EGC) carries cold polar water and sea

ice into the North Atlantic (de Steur et al.,2009). The exported freshwater and sea ice through

EGC comprise roughly half of the total freshwater ﬂux from the Arctic Ocean (Serreze et al.,

2006). While the freshwater outﬂow through Fram Strait may vary from year to year (due to

an alternative exit through Bering Strait;Rabe et al.,2009), almost the entire sea-ice ﬂux from

the Arctic Ocean is exported by EGC (Kwok,2009). The area of the exported sea ice increases

with a trend of 10% per decade since 1990 (Renner et al.,2014). However, this positive trend is

compensated by continuous thinning of the sea ice, and the total annually exported volume of

sea ice does not show a signiﬁcant increase (Zamani et al.,2018). These observations suggest

that the exported sea ice through the Fram Strait is changing, from a thicker old sea ice to thinner and younger sea ice.

The deep-sea LTER observatory HAUSGARTEN was established in 1999 by the Alfred Wegener Institute, Helmholtz-Center for Polar and Marine Research (AWI). It is located in a highly produc-tive transitional area between sea-ice covered and sea-ice free regimes of the Fram Strait (i.e.,

marginal ice zone;Smith Jr et al.,1987). The observatory consists of 21 permanent stations

which are sampled repeatedly in annual summer expeditions since 1999, and includes moni-toring the diversity of all faunal size classes, as well as, biogeochemical measurements (e.g., chlorophyll a and nutrient concentrations). The observations are complemented by

(26)

continu-ous year-round oceanographic measurements collected by autonomcontinu-ous instruments mounted

on moorings (Soltwedel et al.,2005). The observatory covers continental shelves with water

depths of few hundred meters down to the deepest point of the Arctic Ocean - the Molloy Deep, at around 5600 m water depth. The stations array covers both the typically sea-ice covered area in the Arctic outﬂow (EGC), and the seasonally sea-ice covered area in the Atlantic inﬂow (WSC)

(Soltwedel et al.,2005). These distinct conditions between the current systems provide a

valu-able opportunity for studying ecological processes across strong gradients of temperature, and ice cover, in the deep water column and on the shallow continental shelves. Moreover, a compar-ison between the increasing Atlantic inﬂow in the east and the Arctic outﬂow in the west, allows the investigation of ongoing changes in the Arctic marine ecosystem. Throughout the years, the time-series at HAUSGARTEN observatory produced important insights into ecological processes,

and temporal variability (Soltwedel et al.,2016).

A comparison between the observations in the distinct pelagic regimes (EGC and WSC) revealed

higher chlorophyll a concentration in the sea-ice free WSC (Nöthig et al.,2015). There was no

direct correlation identiﬁed between the chlorophyll a and the sea-ice concentration. However, as a result of sea-ice melt and solar radiation, the vertical stratiﬁcation of the surface waters

have shown to promote a higher phytoplankton growth (Cherkasheva et al.,2014). There was

also a clear difference in the phytoplankton community composition between the regimes, with a predominance of diatoms in the EGC and a mixed community of haptophytes, dinoﬂagellates

and diatoms in the WSC (Nöthig et al., 2015;Engel et al., 2017). The heterotrophic

bacte-rial communities also showed strong differences between the regimes, with cell densities in the

range of 104-105cells ml-1in the EGC, and in the range of 106cells ml-1in the WSC. Bacterial

productivity and cell-speciﬁc enzymatic activity showed strong differences between the regions as well, with higher values in the EGC, where the OM was enriched in combined carbohydrates

(Piontek et al.,2014). Showing that bacterial growth and degradation activity in WSC and EGC

are regulated not only by the different physicochemical conditions, but also by the compositional differences in OM. To date, almost nothing was known about the bacterial diversity in the dis-tinct pelagic regimes. The results of this thesis provide the ﬁrst taxonomic comparison of these communities across the Fram Strait.

1.3.2 Ecological alterations in the Fram Strait as a result of a Warm-Water

Anomaly

The oceanographic observations at the HAUSGARTEN observatory have captured an AW warm pulse between 2004 and 2007 (i.e., Warm-Water Anomaly), with temperature anomalies of up

to 1°C along the AW inﬂow pathway (Beszczynska-Moller et al.,2012). The ecological impact

during the anomaly has been visible through the entire marine ecosystem of the Fram Strait.

From a higher chlorophyll a concentrations in surface waters (Cherkasheva et al.,2014), over

a lower vertical export of POM (Lalande et al.,2013), to a sharp change in the phytodetritus

concentrations at the seaﬂoor (Soltwedel et al.,2016). While some of the monitored variables

of the ecosystem returned to their previous state (e.g., benthic bacterial communities; Jacob,

2014), for others the original conditions have not been restored. One of the major changes

in the pelagic ecosystem, which has remained after the end of the anomaly, is a shift from a

(27)

1.3. OBSERVING CHANGES IN THE ARCTIC OCEAN 13

Nöthig et al., 2015; Lasternas and Agustí, 2010; Engel et al.,2017). Aggregates formed by

small ﬂagellates, such as the haptophyte Phaeocystis spp., are more buoyant and sink slower in comparison to diatom aggregates. The longer retainment of the aggregates in the surface ocean

allows stronger recycling in the upper water column (Lalande et al.,2013).

Such a fundamental change in the nature of the OM and its vertical distribution, may have a

strong impact on future carbon cycling processes, from the surface down to the seaﬂoor (

Ver-net et al.,2017). However, it is important to note that natural temporal variations of marine

ecosystems may occur across a wide range of timescales from diurnal to decadal dynamics (e.g.,

Gilbert et al.,2012;Fuhrman et al.,2015). Moreover, they may be subject to even longer global

cycles, such as the El Niño-Southern Oscillation or the North Atlantic Oscillation (Ikeda,1990;

Stenseth et al.,2003). Thus, it cannot be concluded with absolute conﬁdence that the observed

alterations in the Fram Strait are a result of global climate change. However, they provide an important insight into the potential future of the Arctic marine ecosystem as a result of further warming of the Arctic Ocean.

1.3.3 Ocean Observing System FRAM

The extensive knowledge acquired from 15 years of observations in the Fram Strait, strengthened the urgent need for integrative, and interdisciplinary, observations not only in the Fram Strait but also in the central Arctic Ocean. In order to do so, in 2014, the AWI and partner institutes in Europe, established the long-term Arctic open-ocean infrastructure project FRAM (FRontiers

in Arctic marine MonitoringSoltwedel et al., 2013). The FRAM Ocean Observing System is

designed according to the extensive knowledge baseline from the HAUSGARTEN observatory,

which provides the starting point for the observing system infrastructure (Figure 1.2). The main

novelty of the FRAM project is the integrated observation of physical (e.g., autonomous

under-water vehicles;Wulff et al., 2016), chemical (e.g., autonomous benthic crawlers;Wenzhoefer

et al.,2016) and biological (e.g., remotely operated vehicles;Katlein et al.,2017) processes, in the water column and at the seaﬂoor, using cutting edge technologies. In order to allow a comprehensive assessment of the ecosystem responses to global change processes in the Arctic Ocean in the next 25 years.

Due to their key role in biogeochemical processes, microorganisms are of central interest within FRAM. In the HAUSGARTEN observatory, the microbial research in the water column has been

focused mainly on eukaryotic biota (Soltwedel et al.,2016), with very little exploratory work on

Bacteria and Archaea. However, in order to better understand natural dynamics of the marine

ecosystem, and have the ability to detect consequences of the environmental changes,

obser-vations of all three domains of life are required (e.g.,Steele et al.,2011). In the framework

of FRAM, all microbial observations are integrated into a Molecular Observatory (MolObs), which aims to conduct standardized molecular-based high-resolution observations of the Arctic Ocean microbiome. The molecular sampling is conducted synchronously with other observato-ries within FRAM (e.g., physical oceanography and biogeochemistry), which provides a unique opportunity for monitoring microbial communities in a comprehensive environmental context. Unfortunately, extensive microbial long-term time-series studies in the framework of oceano-graphic observatories are rare. Unlike other research ﬁelds of oceanooceano-graphic long-term

(28)

obser-vatories (e.g., physical oceanography), microbial oceanography lacks standardization in the ap-plication of high-throughput methodologies. Microbial observations are often dedicated to a specific research questions, and specific methodologies are applied. These discrepancies be-tween studies challenge their comparability, and their incorporation into one large mechanistic system. Thus, one of the challenges currently faced in the FRAM MolObs is the development of a standardized, sustainable, methodological workflow for conducting long-term microbial ob-servations, which is at the same time comparable to other molecular observatories.

1.4 Thesis objectives

The Arctic Ocean is rapidly changing towards a warmer conditions that are altering the entire ecosystem. Long-term time-series, such as the LTER HAUSGARTEN observatory in the Fram Strait, are essential for the detection and understanding of large-scale environmental changes. To date, the microbial research in the water column of the Fram Strait was mainly focused on phytoplankton communities. These primary producers play the key role in the oceanic uptake of

CO2through the ﬁxation of inorganic carbon, and are very important for estimating carbon ﬂuxes

between the atmosphere and the ocean. The heterotrophic bacterial and archaeal communities,

on the other hand, respire a large fraction of the produced organic carbon back to CO2. They

are also involved in other nutrient cycles (e.g., nitrogen cycle) that can both enhance, and limit, the primary production by phytoplankton. Thus, marine microorganisms of all three domains of life are relevant for studying Arctic marine ecosystem and its biogeochemical cycles.

The aim of this thesis was to investigate the composition and diversity, of pelagic bacterial and ar-chaeal communities in the Fram Strait. Addressing their distribution in space, both horizontally and vertically, in relation to environmental and biological parameters. In addition, this work contributed to the establishment of baseline knowledge for long-term microbial observations in the framework of the FRAM project. The ecological objectives of the thesis were: (i) to charac-terize the bacterial and archaeal communities associated with the different water masses of Fram Strait, and (ii) to identify environmental factors, which drive the diversity of these communities. Speciﬁcally the thesis addressed the following questions:

• What are the requirements for long-term microbial observations in the Arctic Ocean? - Time-series microbial observations are of a high relevance for the assessment of the on-going changes, not only in the Arctic ecosystem, but in the entire global ocean. The ob-servations should include various habitats (e.g., water column and seaﬂoor), at various geographic locations, integrated into a holistic evaluation of the ecosystem state. Such a complex task requires an establishment of a network between various long-term micro-bial observatories, and methodological standardization for comparability of biological, and

biogeochemical observations between them. Inchapter 2we reviewed existing microbial

observatories, and suggest potential directions for the establishment of communication and data ﬂows between them.

• Which universal primer set for the 16S rRNA gene should be implemented in

Arc-tic Ocean bacterial observations? - The 16S rRNA gene-based studies of the marine

sediment microbiome, often use the primer set 341F/785R that targets the V3-V4

(29)

1.4. THESIS OBJECTIVES 15

Figure 1.2: (A) Map of the FRAM infrastructure in the Arctic Ocean. The extremely low sea ice extent in September 2012 is shown as a white/blue area. (B) The section map represents the Fram Strait. The HAUSGARTEN observatory stations are shown by red dots, blue arrows indicate the EGC and the red arrows indicate the WSC. Map: AWI/ Laura Hehemann, Ingo Schewe. (C) A section across the Fram Strait at 79°N. The dierent colours represent the distinct water masses in the upper 1000 m of the water column. The Polar Surface Water (PSW), which are carried by the EGC, are originating in the Polar Mixed Layer and the Arctic halocline. The Atlantic Water (AW) are carried by the WSC. The Warm Surface Water (PSWw) are representing the sea-ice melting water, and the Arctic Intermediate Water (AIW) represent the mesopelagic water layer in the Arctic Ocean. The water masses were identied using PHC3.0 (updated from Steele et al., 2001) annual temperature, salinity and density values, based on denitions inRudels et al.(2005). Map: AWI/ Claudia Wekerle.

(30)

pelagic microbiome often implement the primer set 515F-Y/926R that targets the V4-V5 hypervariable regions of the 16S rRNA gene. In the framework of the FRAM MolObs, we aim to integrate microbial observations from the different habitats (water column and sed-iment) into one holistic ecosystem observation. Such integration requires methodological standardization between the sampling and analysis procedures of the different habitats.

The methodological comparison of the different primer sets inchapter 3, will allow the

se-lection of a single primer set for future standardized observations of bacterial communities in the Arctic marine ecosystem.

• Which environmental parameters shape summer pelagic bacterial diversity in the

photic layer of the Fram Strait? - Temperature has been identiﬁed as the main driver of

the microbial diversity in the surface waters of the global ocean. It has also been shown that light driven phytoplankton blooms have a strong impact on the associated bacterial diversity. In the Fram Strait, there is a strong temperature gradient across the EGC and

WSC waters, ranging between -1°C and +8°C, respectively. Due to the distinct sea-ice

conditions this temperature gradient is also coupled with strong differences in light pene-tration through the water column, which may have an impact on the phytoplankton bloom. This, suggests that bacterial diversity in the Fram Strait might be driven by both tempera-ture and differences in sea ice cover. However, previous observations of bacterial activity in the Fram Strait revealed stronger correlation with phytoplankton bloom conditions, rather than physical characteristics of the water masses. Thus, the bacterial community is expected to exhibit stronger dissimilarity across the Fram Strait as a result of different sea-ice regimes, associated with different phytoplankton bloom conditions. Addressed in

chapter 4.

• How do sea-ice conditions affect the vertical connectivity between surface and deep

ocean microbial communities? - The OM produced by phytoplankton is exported to the

deep ocean through sinking aggregates. It has been shown that these sinking aggregates may provide a connectivity vector between surface and deep ocean microbial communities. Sea ice plays a key role in regulating the primary production in the Arctic marine ecosys-tem. Furthermore, sea-ice covered and sea-ice free waters are characterized by different phytoplankton communities, which have different sinking velocities to the deep ocean. Thus, fast sinking diatom aggregates in sea-ice covered regions are expected to provide a stronger vertical connectivity between surface and deep ocean microbial communities.

Addressed inchapter 5.

1.5 Publication outline

Chapter

2 -

Marine Microbes in 4D – Using time series observation to assess

the dynamics of the ocean microbiome and its links to ocean health

Pier Luigi Buttigieg, Eduard Fadeev, Christina Bienhold, Laura Hehemann, Pierre Offre and Antje Boetius

Ecological observations of pelagic bacterial and archaeal communities in the Atlantic-Arctic boundary zone